Subjective tinnitus is perception of sound like ringing, buzzing, or hissing in the absence of external stimulation. Tinnitus is a significant neurological disorder that affects 51 million US citizens. There is no cure. Twelve million patients seek treatment for disturbances of sleep, affect and concentration and 3 million are unable to function occupationally. Current treatments do little to actually reduce tinnitus perception. In contrast, repetitive transcranial magnetic stimulation (rTMS) is a non-invasive method of regional brain stimulation that can significantly reduce subjective tinnitus in 50% of patients; however, the treatment effect is temporary, lasting 1 to 2 weeks after a week-long course of standard treatment. Our preliminary work indicates that this limitation can be overcome by changing the way rTMS is applied. We have observed that the rTMS effect on tinnitus is very reliable among persons who respond positively to a standard course of treatment (i.e., treatment responders) and that reapplying rTMS when tinnitus returns (i.e., maintenance treatment) produces added benefits that are sustained over time. We hypothesize that maintenance rTMS works by reversing pathological processes which are set in motion by a loss of normal sensory input (i.e., a thalamocortical dysrhythmia) and we will test this hypothesis using high density EEG recordings before and after treatment. We expect that maintenance treatment will improve tinnitus in an additive fashion and that it will impede relapse relative to standard treatment and to a sham-treatment control. We will focus the study on 30 treatment responders who will receive sham and active maintenance rTMS using a blinded study design with treatment crossover. Thirty non responders will be studied to learn what patient characteristics predict a treatment response. Maintenance treatment will consist of four, 3-day courses of rTMS separated by fixed 3-week intervals. Primary outcome measures will include a questionnaire of tinnitus severity, visual analogue ratings of tinnitus loudness and annoyance, forced- choice measures of improvement/worsening, a physically anchored measures of tinnitus frequency and loudness and the spectral power and coherence of EEG frequency bands. The timing and duration of maintenance treatment effects will be assessed by counting the number of days between the end of a course of treatment and the return of tinnitus. Follow-up assessment will be conducted at 3 months to evaluate change in the primary outcome measures and in the baseline assessments of clinical, behavioral, emotional, and perceptual aspects of tinnitus related to quality of life. Analyses will compare change scores from baseline on the primary outcome measures between standard and active maintenance treatment and between active and sham maintenance treatment. Attrition and missing data will be treated with subject replacement. This project is significant because it has potential to produce a treatment that decreases tinnitus perception chronically. Our innovative approach of using maintenance rTMS can shift current paradigms and models of how rTMS is applied in other clinical disorders treated with rTMS. We expect maintenance rTMS to benefit half of patients with tinnitus. We anticipate that these findings will support larger, multisite trials that can establish rTMS as a frst line treatment approach for tinnitus and gain FDA approval for this use.